Cytomegalovirus (CMV) pneumonia continues to be the major infectious cause of mortality following allogeneic marrow transplant. Although antiviral therapy has been shown to reduce the virus titer in the lung, mortality is not reduced. The study of local pulmonary immunity may ultimately provide a rationale for new therapeutic strategies that could be applied, perhaps locally to the lung, for the control of this severe infection. There is increasing evidence that immune responses mediated by pulmonary cells may be distinctly different from those mediated by peripheral blood cells. Bronchoalveolar lavage has emerged as a useful technique for obtaining representative cells from the lung for characterization of pulmonary immune responses. Evaluation of cytotoxic responses by natural killer cells derived specifically from the lungs of marrow transplant recipients may provide an explanation for: 1) the occurrence of CMV pneumonia after marrow transplant, and 2) the progression of CMV pneumonia despite antiviral treatment. The proposed research will determine the presence and level of cytotoxic activity in the lung and will compare the relative contribution of natural killer cells, cytotoxic T lymphocytes, and alveolar macrophages to lysis of CMV-infected and uninfected fibroblast target cells. The mechanisms which influence natural killer cell activity will be evaluated, focusing primarily on the regulatory role of alveolar macrophages and the production of monokines, including tumor necrosis factor. In addition the role of local interleukin 2 and interferon production in the maintenance or enhancement of natural killer cell activity will be evaluated. Whether differences in regulation of natural killer cell activity in the lung can explain differences observed between responses by pulmonary and peripheral cells and between responses by cells from normal subjects and marrow transplant patients will be examined. Finally, responses by pulmonary and peripheral cells from marrow transplant patients with and without CMV pneumonia and from normal subjects will be compared to determine if differences in activity can explain the occurrence and outcome of CMV pneumonia after marrow transplant.